Functional Specificity Among Hsp70 Molecular Chaperones

See allHide authors and affiliations

Science  17 Jan 1997:
Vol. 275, Issue 5298, pp. 387-389
DOI: 10.1126/science.275.5298.387


Molecular chaperones of the 70-kilodalton heat shock protein (Hsp70) class bind to partially unfolded polypeptide substrates and participate in a wide variety of cellular processes. Differences in peptide-binding specificity among Hsp70s have led to the hypothesis that peptide binding determines specific Hsp70 functions. Protein domains were identified that were required for two separate functions of a yeast Hsp70 family. The peptide-binding domain was not required for either of these specific Hsp70 functions, which suggests that peptide-binding specificity plays little or no role in determining Hsp70 functions in vivo.

Hsp70 proteins function in a diverse set of processes, including protein folding, multimer association and dissociation, translocation of proteins across membranes, and regulation of the heat shock response (1). All eukaryotic cells use multiple Hsp70s to carry out these functions; in the yeast Saccharomyces cerevisiae, 14 Hsp70s are divided into at least five functionally distinct families. Hsp70s from different families are highly conserved but cannot function interchangeably (2, 3); the basis of this functional specificity is not well understood. Each Hsp70 consists of a highly conserved NH2-terminal 44-kD adenosine triphosphatase (ATPase) domain, a less well conserved 18-kD peptide-binding domain, and a COOH-terminal 10-kD variable domain of unknown function (4, 5, 6, 7, 8). Because differences in peptide-binding specificity have been identified among Hsp70s (9, 10, 11), it has been hypothesized that peptide binding plays a central role in determining the functional specificity of each Hsp70. Here, we challenge this hypothesis by determining the source of functional differences between two families of yeast Hsp70s.

The Ssa and Ssb Hsp70 families of S. cerevisiae share 60% amino acid identity and reside in the cytosol (2) but have distinct, nonoverlapping functions in vivo (12). To test the role of different Hsp70 domains in determining functional specificity, we generated gene fusions that encoded chimeric proteins containing all combinations of the ATPase, peptide-binding, and variable domains of Ssa1 and Ssb1 (13). Chimeric junctions were placed precisely between the three Hsp70 domains (5, 6, 7, 8, 11).

To directly test the hypothesis that functional specificity is determined by peptide binding, we investigated whether a chimera containing the Ssa1 peptide-binding domain could rescue two phenotypes specific to the Ssb family of Hsp70s: cold sensitivity and hygromycin B sensitivity. The chimeric protein BAB (14), in which the peptide-binding domain of Ssb1 has been replaced by that of Ssa1, was able to rescue both Ssb-specific phenotypes (Fig. 1). Thus, the functional differences between the Ssa and Ssb protein families were not determined by the peptide-binding domain. Although the peptide-binding domain contains all of the residues that contact the peptide substrate (8), the variable domain was also included in most previous studies of peptide-binding function (6, 9, 11, 15) and thus might modulate peptide binding. Because the variable domains of Ssa1 and Ssb1 are only 14% identical, they are a potential source of functional differences. However, the chimera BBA, containing the variable domain of Ssa1, was able to rescue both cold sensitivity and hygromycin B sensitivity (Fig. 1); hence, the variable domain was not required for the rescue of either Ssb-specific phenotype.

Fig. 1.

Rescue of Δssb1 Δssb2 growth phenotypes by chimeric Hsp70 proteins. Δssb1 Δssb2 cells containing each chimera on a centromeric plasmid were grown on SC-ura media at 18°C for 8 days, or at 30°C in the presence or absence of hygromycin B (70 μg/ml) for 4 days. Immunoblotting experiments demonstrated that the wild-type proteins and all chimeras were stably expressed at similar levels at both 18° and 30°C (16).

Because neither the peptide-binding nor the variable domain was required for the rescue of Ssb-specific phenotypes, we examined all of the Ssa1-Ssb1 chimeras to determine the source of Ssb-specific function (Fig. 1). Both Δssb1 Δssb2 phenotypes were rescued by several chimeric proteins; however, the two phenotypes were not rescued by the same chimeras. The cold-sensitive phenotype was rescued by all of the chimeras that contained the Ssb1 ATPase domain, but not by those that contained the Ssa1 ATPase domain (Fig. 1). Thus, the ATPase domain of Ssb1 was the only domain specifically required for the rescue of cold sensitivity.

The rescue of hygromycin B sensitivity was more complex. No single domain of Ssb1 was necessary for the rescue of this phenotype; BBA, BAB, and ABB all conferred resistance to hygromycin B (Fig. 1). Moreover, no single domain was sufficient; cells containing AAB, ABA, or BAA all remained sensitive to hygromycin B. Rather, the different domains of Ssb1 functioned additively, such that any two could confer resistance to hygromycin B. These domains could not function cooperatively when present in different chimeric proteins, however; when any two of the chimeras BAA, AAB, and ABA were coexpressed, no rescue of hygromycin B sensitivity was observed (16).

Ssb proteins appear to play a role in translation since Δssb1 Δssb2 mutants are sensitive to hygromycin B, and up to 70% of Ssb protein is associated with translating ribosomes (17). We examined five chimeras (18) to determine whether the rescue of hygromycin B sensitivity was correlated with polysome association. Polysomes were separated from free ribosomes by centrifugation through a sucrose gradient (Fig. 2A). Both Ssa and Ssb proteins were present in high-molecular weight complexes that were distributed throughout a sucrose gradient (Fig. 2B). These complexes were distinguished by treating extracts with ribonuclease (RNase) A, which disrupts the polysomes. Ribosome-associated proteins such as Ssb were shifted toward the top of the gradient (17), whereas Ssa was unaffected.

Fig. 2.

Association of chimeric Hsp70 proteins with polysomes. (A) Separation of free ribosomes (80S peak) and polysome complexes on a sucrose gradient (25). RNase A treatment disrupts polysome complexes, causing a shift into the 80S peak. (B) Immunoblots of polysome profile fractions. Each panel shows fractions taken from across a sucrose gradient, with the positions of the 80S peak and the polysomes indicated. Immunoreactive bands are identified in the top panel of each pair. Compare each upper panel (−RNase A) with its partner below (+RNase A). Proteins associated with polysomes are shifted toward the 80S peak upon RNase A treatment.

Sucrose gradients were used to examine the polysome association of the chimeras. Chimeras that failed to rescue hygromycin B sensitivity (BAA and AAB; Fig. 1) also failed to associate with polysomes (Fig. 2B). In contrast, ABB and BAB, which conferred hygromycin B resistance, associated with polysomes. Thus, neither the ATPase domain nor the peptide-binding domain was required for the Ssb-specific association with translating ribosomes. Moreover, for all of the chimeras except BBA (19), there was a correlation between the rescue of hygromycin B sensitivity and an association with translating ribosomes, which suggested that both are the result of a single ribosomal function.

In contrast, the cold-sensitive and hygromycin B-sensitive phenotypes of a Δssb1 Δssb2 strain were separable (Fig. 1). The chimera BAA rescued only the cold-sensitive phenotype of Δssb1 Δssb2 cells, whereas ABB rescued only hygromycin B sensitivity. This separation of phenotypes suggested that cold sensitivity and hygromycin B sensitivity represent two distinct cellular functions of the Ssb proteins. The chimera BAB rescued both of these phenotypes; thus, the peptide-binding domain of Ssb1 was dispensable for two different Ssb-specific functions. In addition, the ability of BAA to rescue cold sensitivity in the absence of detectable polysome association suggested that Ssb proteins performed these two functions at different sites. This suggestion was supported by cotransformation experiments. When both BAA and ABB were present in the same cell, they did not interfere with one another but functioned additively, rescuing both cold sensitivity and hygromycin B sensitivity (16, 20).

Together, these results indicate that functional specificity among the Hsp70 class of molecular chaperones does not depend on peptide-binding specificity. We favor a model in which functional specificity is determined by physical interactions between one or more Hsp70 domains and other components of the cellular machinery (21). Once established, these interactions may be stabilized by binding to substrate polypeptides; however, we view peptide binding as an activity of the Hsp70s, similar to their ATPase activity, that is ferried to particular functions by the specific interactions of other parts of the protein. Such interactions may serve to target Hsp70 to a site of function, such as the ribosome, or to direct association with cohort proteins such as DnaJ homologs, which function cooperatively with Hsp70s (22, 23).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
View Abstract

Navigate This Article